A Similarity Link Prediction Method in Complex Network Based on Endpoint Clustering

摘要

Link prediction aims to predict the probability of the existence of links between two endpoints in complex network. Many methods ignore the clustering of endpoints when calculate the similarity between two endpoints. To distinguish the contribution of endpoints clustering, we propose a similarity link prediction method based on endpoint clustering. In order to improve the link prediction accuracy, the method considers both the common neighbor and endpoint clustering. Empirical study on six real networks has shown that the method we proposed can achieve a good performance, compared with CN, AA, RA, LP and Katz.